The primary objective of this study is to identify the diagnostic features of autism that are present in children under 36 months of age. Although many authors have alluded to developmental changes in symptom expression, empirical study of the early manifestations of autism has been scant. The study of early symptoms has important clinical and theoretical implications. At a clinical level, children identified at young ages can accrue the benefits of early participation in specialized intervention programs. Moreover, identification of early patterns of symptom expression can contribute to our understanding of the primary deficits of autism. This study will employ a prospective, longitudinal research design. Fifty children under 36 months of age who demonstrate behavioral characteristics within the "autism spectrum" will be followed through the age of 4 years, at which time they will receive a definitive clinical diagnosis. Clinical and diagnostic assessments will occur at the ages of 2, 3, and 4 years, and will include measures of developmental/cognitive functioning, receptive and expressive language skills, adaptive behavior, and severity of autism. Diagnostic assessment will be performed independently by two experienced clinicians. Home and classroom observations will be conducted at the ages of 2 and 3 years to evaluate the influence of environmental factors on the early developmental course of this disorder. The design and methodology of this study will enable us to determine: 1) the reliability of the diagnosis of autism made below the age of 36 months; 2) the early symptoms that are predictive of a definitive diagnosis of autism at 4 years; and 3) the factors that are associated with changes in symptomatology over the two-year period. This project represents an improvement over previous studies in its inclusion of a larger sample, use of prospective rather tan retrospective methodology, evaluation of interobserver agreement for clinical diagnosis, and consideration of environmental influences on changes in symptomatology.